Hardik Jindal

Optimising LLM inference @scorelabsai | GSoC'26,24 @c2si | EE'26 @iitkanpur

[Hardik Jindal @Hads1408]

@CerebralCoding_ All of these techniques fail at high context/ high batch size. Wonder how this performs..

[Hardik Jindal @Hads1408]

@shafu0x Hope you changed it all occurences 😅

[Hardik Jindal @Hads1408]

@OGKAIM Hopefully it'll be a dense model

[Hardik Jindal @Hads1408]

@chenlin_meng @scorelabsai 💪

[Hardik Jindal @Hads1408]

@lasse_ai Any blog about this technique?

[Hardik Jindal @Hads1408]

@thsottiaux There will come a time when no human might understand how shit works, and that will be another problem for humanity then..

[Hardik Jindal @Hads1408]

@coffeecup2020 whereeeeee??

[Hardik Jindal @Hads1408]

@tonymongkolsmai Does it support h100/200?

[Hardik Jindal @Hads1408]

@alokbishoyi97 Don't see anything mate :(

[Hardik Jindal @Hads1408]

@TrijalSrivasta3 looks impressive. thanks!

[Hardik Jindal @Hads1408]

This is simply badass!

[Liquid Energy (YC F26) @_LiquidEnergy]

We built a new type of datacenter for AI inference. We call it NanoDC.

NanoDC is a solar-powered autonomous datacenter that can be deployed anywhere on the planet in 4 weeks.

11 Trillion will be invested in AI compute over the next 4 years. Most of this will be for inference and most of it will be severely delayed: stuck in grid interconnect queues, site permits, or supply chain issues. NanoDC sidesteps these issues completely.

NanoDC is built rapidly at a factory and shipped to the site ready to be turned on in 30 minutes. It generates its own power with solar and batteries. You open the door and the solar panels fold-out like an accordion. The compute is cooled with patent-pending high-density air cooling. It connects to the internet via Starlink.

No construction. No grid interconnect. No Water supply.

It never waits for transformers and turbine generators either.

Getting more compute is as easy as dropping more units and turning them on.

The first units will be live this month. Thousands will be shipping starting 2027.

Sunlight in. Tokens out.

[Hardik Jindal @Hads1408]

@ivanfioravanti Maybe they are still thinking like the qwen 3 😂

[Hardik Jindal @Hads1408]

@QwenAIFans I see. How was support for qwen3.8 27b added for this? Did it hurt/improve its overall performance?

[Hardik Jindal @Hads1408]

@rikiyatakehi @a1zhang @lateinteraction @jayden_teoh_

[Hardik Jindal @Hads1408]

@morqon @CheekyPint Yeah low key agree with Sam here. I think less than 1% of the people will be using local AI

[Hardik Jindal @Hads1408]

@Deaphen Check on your family man why are you posting here just after 45 seconds 😂

[Hardik Jindal @Hads1408]

@roy_eisenstadt Comparision against EAGLE4?? I doubt it can beat it.

[Hardik Jindal @Hads1408]

Impressive. This is better than Qwen3.8-27b across all benchmarks. Would definitely try this one!

[Nanbeige @nanbeige]

🚀Excited to announce Nanbeige4.3-33B!
To push the performance ceiling of small agents, our strategy in Nanbeige4.3 starts with building a powerful teacher first.
Nanbeige4.3-33B delivers SOTA agentic performance among open models at ~30B parameters:
* Strong results on agentic coding and cowork benchmarks.
* When integrated into mainstream agent harnesses, it tackles real-world tasks with cost-efficient inference.
* It will serve as a teacher for upcoming smaller Nanbeige models, starting with Nanbeige4.3-3B.

https://t.co/tSToQgZf2O
https://t.co/D228euvybz

#LLM #Agent